Fastest
Eureka on the web
When you want the answer in the next five minutes.
The agent works the prompt against patents and technical literature, citing every source.
Run your analysis now →https://www.patsnap.com/resources/blog/rd-blog/ranking-models-and-feature-interaction-patent-landscape/ · Patsnap · data cut-off 2026-07-31 · downloaded from the live page
Patent Landscape · Recommender SystemsRanking Models and Feature Interaction Patents: who holds the claim space
- Concentrated at the top. The leading assignee alone holds 38 of 221 records, and the top five together account for 44.8% of all records in scope.
- Filing has cooled from its peak. Activity rose to 44 records in 2022 before easing to 22 by 2024, a -12% change over that span, though the newest years are still filling in.
- Core computing classes dominate. G06F and G06N each cover more than half of the 221 records, while image/video and transmission-layer combinations remain in single digits.
221
Published Records
45%
Top-5 Share of All Records
-12%
Filing Growth 2021→2024
US
Leading Jurisdiction
Filing growth compares 2021 (25 records) with 2024 (22) — a three-year span. 2024 is the most recent year we treat as complete: publication lags filing by roughly 18 months, so 2025 onwards are still filling in and any growth rate that ends there would understate the field. Top-5 share is the combined record count of the five largest assignees divided by all 221 records in scope (CR5), not by the ranked leaders only.
Published byPatsnap Research··7 min readSourced from Patsnap Eureka
Field Overview
What this landscape covers
Ranking models and feature interaction sit at the core of modern recommendation and ad-serving systems, where click-through-rate estimation depends on how features are crossed, embedded and calibrated before a ranking decision is made. This landscape covers 221 published records filed between 2015 and mid-2026 that combine ranking-model or feature-interaction terminology with technical mechanisms such as embedding tables, position-bias correction, delayed feedback and multi-task objectives.
Filing activity spans search advertising, e-commerce ranking and general recommender infrastructure, with the heaviest concentration in digital-processing and AI-model classifications rather than narrower business-process claims. The picture below covers assignee concentration, filing trend, technology composition and the most-cited records in the set.
Records by publication year, 2017-2026
Source: Patsnap Eureka. Assignee ranking and totals. Derived from a Patsnap search on Ranking Models and Feature Interaction covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Let an AI agent run this analysis on your own technology
Pick a task. Every answer cites the patents behind it.
10,000 free credits to start
Filing Data
How filing activity has moved since 2017
Publication counts across the 221 records in scope show where activity concentrated and where it has since levelled off. Because publication lags filing by roughly 18 months, the final one or two years on this chart will keep filling in as more applications publish.
Filing trend, 2017-2026
Annual filings rose from 4 in 2017 to a peak of 44 in 2022, then eased to 22 by 2024 — a -12% change over the 2021-2024 span. 2025 and 2026 figures are still incomplete due to publication lag and should not be read as a decline.
Technology composition by IPC subclass
G06F (electric digital data processing) and G06N (AI-based computing) each cover more than half of the 221 records, confirming that most filings frame ranking and feature-interaction work as core computing or model architecture rather than as a narrower business-process claim under G06Q.
Shares are the percentage of the 221 records in scope. A patent can carry several IPC classes, so the shares add up to more than 100%.
Source: Patsnap Eureka. Filing trend and technology composition. Derived from a Patsnap search on Ranking Models and Feature Interaction covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.
Go deeper on Ranking Models and Feature Interaction with Eureka
This page is one run against one query. Ask Eureka your own question about ranking models and feature interaction and every answer comes back with the patent numbers behind it.
Try EurekaReference Filing
A representative early claim in this space
Representative Record
Effective ad placement (US20110258033A1)
MICROSOFT TECHNOLOGY LICENSING, LLC
Search results and ads that satisfy a query may be formatted into a search results page where an initial placement of ads may result in an estimated click through rate for the presented ads. An adjustment factor may be determined based on parameters associated with the search results, such as the clicks on the search results themselves, and applied to determine if an estimated click through rate will change with respect to position when there are differing numbers of mainline and sidebar ads.Filed by Microsoft Technology Licensing, dated 2011-10-20 — one of the earlier filings in this dataset tying click-through-rate estimation to ad placement position.
Most-cited records in this dataset
| # | Publication no. | Patent title | Citations |
|---|---|---|---|
| 1 | US20150317310A1 | Method and system for evaluating query suggestions quality | 251 |
| 2 | US20170308609A1 | Multi-result ranking exploration | 143 |
| 3 | US20180137404A1 | Joint learning of local and global features for entity linking via neural networks | 109 |
| 4 | US20120303615A1 | Image-based popularity prediction | 91 |
| 5 | US20130179252A1 | Method or system for content recommendations | 78 |
| 6 | US20260046317A1 | AI-Optimized Memory Fabric for Large Contexts and Multimodal Workloads | 60 |
| 7 | US8037043B2 | Information retrieval system | 52 |
| 8 | US20100153370A1 | System of ranking search results based on query specific position bias | 46 |
| 9 | US20160321716A1 | System, method, and non-transitory computer-readable storage media for enhancing online product search throug… | 44 |
| 10 | US20100241597A1 | Dynamic estimation of the popularity of web content | 43 |
Citation counts favour older filings in any searched corpus — treat them as a signal of influence within this dataset, not as a ranking of current technical importance.
Each row carries its publication number; clicking a row searches Eureka by that number.
Source: Patsnap Eureka. Citation counts and representative records. Derived from a Patsnap search on Ranking Models and Feature Interaction covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Run it yourself
Put your own technology through the same analysis
Where to run it
For builders
MCP server & REST API
When it has to run inside your own pipeline.
Patent search, landscape analysis and assignee resolution as MCP tools. Drop them into any agent framework, or call REST directly.
Browse MCP servers →Signal Check
What the filing pattern tells a decision-maker
These figures come directly from the 221 records in scope and point to where claim space is occupied, where it has room, and which years are still understated by publication lag.
Concentration
44.8%
of 221 records held by top 5
A small group sets the terms
The top five assignees combined account for 44.8% of all 221 records in scope, with the leader alone holding 38. That leaves a long tail across the remaining 63 of the 68 ranked companies, each with a modest filing count.
Ranked across 68 companies
Momentum
-12%
change, 2021 to 2024
Peak filing has passed, for now
Filings peaked at 44 in 2022 before easing to 22 by 2024. Because publication lags filing by roughly 18 months, 2025-2026 counts are not yet complete and should not be read as continued decline.
Peak year: 2022 at 44 records
Composition
62.4%
of records classified under G06F
Core computing framing dominates
G06F and G06N together cover the large majority of the 221 records, confirming most filers frame their claims as general digital processing or AI-model architecture rather than a narrower business-process angle.
G06N covers 59.3% separately
Eureka AI Agent
Looking for what nobody has claimed yet?
Eureka can read the same corpus for gaps instead of for coverage: under-claimed branches adjacent to ranking models and feature interaction, with the prior art for and against each one.
Source: Patsnap Eureka. Co-assignee relationships and derived observations. Derived from a Patsnap search on Ranking Models and Feature Interaction covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Competitive Landscape
Who is filing, and where the field is still open
Filing activity in this dataset is led by a small group of large technology and commerce companies, with a long tail of single- or low-count filers behind them. Momentum in the most recent year has cooled across several of the leading names, which is consistent with the broader 2021-2024 trend rather than a sign any one company has exited the space.
Leader
38 records
held by the top-ranked assignee
A single company sets the pace
The leading assignee's 38 records are more than three times the fifth-place count of 11, marking a steep drop-off typical of a field where one or two large filers built early portfolios in CTR estimation and feature crossing.
Fifth place holds 11 records
Mid-field
11 to 5
records, fifth to tenth place
A competitive middle tier
Between fifth and tenth place, record counts fall from 11 to 5, indicating a group of companies with sustained but narrower filing programmes rather than one-off applications.
Top 10 combined: 59.3% of 221 records
Long tail
68 companies
in the ranked field
Most filers hold a handful of records
With 68 companies ranked and the top ten already covering 59.3% of all 221 records, the remaining filers each hold small counts — useful for spotting niche technical angles rather than broad portfolios.
Long tail below rank 10
Recent-year filing momentum by assignee
| Assignee | Recent year | YoY |
|---|---|---|
| Microsoft Technology Licensing, LLC | 0 | -100% |
| Tencent Technology (Shenzhen) Co., Ltd. | 0 | -100% |
| Huawei Technologies Co., Ltd. | 0 | — |
| Samsung Electronics Co., Ltd. | 0 | -100% |
| eBay Inc. | 0 | — |
| Walmart Apollo, LLC | 0 | — |
| Microsoft Corporation | 0 | — |
| 0 | — |
Source: Patsnap Eureka. Assignee-level momentum. Derived from a Patsnap search on Ranking Models and Feature Interaction covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Next Steps
Where to take this analysis next
The figures above establish the shape of the field. The next layer of work is claim-level: reading granted claims against a specific product design, not just counting records.
Check freedom to operate against the top filers
With 44.8% of the 221 records held by five assignees, a claim chart against those portfolios specifically is a faster first filter than reviewing the full ranked field.
Run a claim comparison in EurekaTrack the under-claimed combinations before they fill in
Position-bias and delayed-feedback claims tied to visual or transport-layer signals currently sit in single-digit IPC shares — worth monitoring before filing volume catches up.
Set up monitoring in EurekaRe-check the 2025-2026 trend once it publishes fully
Because publication lags filing by roughly 18 months, the current dip after the 2022 peak is not yet a confirmed slowdown and should be revisited as more records publish.
Explore the filing trend in EurekaSource: Patsnap Eureka. Forward-looking reading of the same dataset. Derived from a Patsnap search on Ranking Models and Feature Interaction covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
FAQ
Common questions about this landscape
Filing is concentrated at the top: the leading assignee alone accounts for 38 of the 221 records in scope, and the top five combined account for 44.8% of all records. The tenth-ranked assignee holds only 5 records, which shows a fast drop-off from the leader to the rest of a 68-company ranked field. This concentration means a small number of portfolios shape the core claim space around CTR estimation and feature crossing, while the remaining companies each hold a handful of filings in narrower niches.
Filings rose steadily from 4 in 2017 to a peak of 44 in 2022, then eased to 22 by 2024, a -12% change over that three-year span. That is a real signal for the most recent complete years, but 2025 and 2026 figures are still incomplete because publication typically lags filing by about 18 months. Treat the 2022 peak as the high point so far rather than assuming the field is now shrinking.
G06F (electric digital data processing) covers 62.4% of the 221 records and G06N (AI-based computing) covers 59.3%, making these the two dominant classification areas. G06Q (business and commerce data processing) follows at 30.3%, tying many filings to specific commercial contexts like advertising or e-commerce ranking. Smaller classes such as G06V (image/video recognition) and H04L (digital transmission) sit in single digits, marking areas with comparatively less claim density.
The subclasses with the lowest share of the 221 records — G06V at 8.1%, G06K and G06T at 4.1% each, and H04L at 2.7% — carry the least filing density in this dataset. Combinations that tie ranking or feature-interaction claims to image/video signals or transport-layer mechanics are comparatively under-claimed relative to the dominant CTR-and-embedding combination. Jurisdictionally, India's 4 filings against 126 in the United States also suggest room for filers who have already committed to a technical route.
US20110258033A1, assigned to Microsoft Technology Licensing, covers adjusting an ad's estimated click-through rate through a placement-based adjustment factor, including a mainline-versus-sidebar exclusivity mechanism. Its scope is tied specifically to that placement-adjustment approach, not to CTR estimation or feature interaction generally, so systems using different signals or that don't adjust placement position likely sit outside it. Anyone building a click-adjustment or ad-placement ranking feature should still review its granted claims directly rather than relying on the abstract alone.
Answers are grounded in the same dataset. Derived from a Patsnap search on Ranking Models and Feature Interaction covering 2015–2026, data cut-off 2026-07-31. Counts reflect published records only and shift as new filings publish.Run this in Eureka MCP
Research Ranking Models and Feature Interaction in depth with Eureka
Go past this page: query the whole ranking models and feature interaction corpus yourself, in your own scope.
Every answer comes back with patent numbers you can open.
Disclaimer. This page is generated from Patsnap Eureka data drawn from a limited snapshot of global patent and scientific-literature records, and is provided for general information and reference only.
Patent data carries inherent limitations: recent filings (typically the most recent 18–24 months) are under-counted due to standard publication lag; counts may be reported at either a patent-family or a patent-record basis and are not always directly comparable; classification, applicant-name, and citation data may contain errors, duplicates, or omissions; and the underlying search query defines and constrains the scope shown. As a result, the analysis may be incomplete or inaccurate and may not reflect the full technology landscape.
Nothing on this page constitutes an exhaustive prior-art, novelty, freedom-to-operate, or validity search, nor does it constitute legal, financial, investment, or professional advice, and it should not be relied upon as such. Any patent, commercial, or strategic decision should be verified independently and reviewed with qualified patent, legal, and domain professionals. Patsnap makes no warranties, express or implied, as to the accuracy, completeness, or fitness for any particular purpose of the information presented.
Machine translation. Assignee and organisation names originally recorded in Chinese, Japanese or Korean have been rendered into English by an AI translation step so that the tables stay readable. These renderings are best-effort and may not match a company's registered English name; the original name is what the underlying patent record carries, and it is what any Eureka query launched from this page uses.